Use YAML interface to myApp

[stsievert]

Use YAML interface to myApp

[stsievert]

Passing in a file as a string doesn't address my concern. Each file has a specific format: each line has a (string, string, int) or something. We have a system to pass in arguments to the app. Use it.

I don't know if this will effect the speed issues. It will certainly help diminish complication.

[ayonsn017]

Different algorithms take as input different types of files. The format within the file are different. Hence I cannot specify it in yaml in a single format.

[stsievert]

Looking in local/data I see two file types: one JSON object and many different CSVs with (almost) the same header. I'm assuming the CSV files are the different files? This can be worked around.

In the YAML file you can specify a list of dicts. In each dict you can specify the keys. Each key can be optional, which is what I think you're concerned about.

You'd want to pass something like

[{'mol1': 'HCN', 'rep1': 'SF', 'mol2': 'HCN', 'rep2': 'SF', 'same': True},
 {'mol1': 'H20', 'rep1': 'SF', 'mol2': 'H20', 'rep2': 'SF', 'same': True}
 ...
]

where each dictionary could optionally include a probability. This could be done by a YAML snippet like

questions_to_ask:
  type: list
  values:
    type: dict
    values:
      mol1:
        type: str
      rep1:
        type: str
      mol2:
        type: str
      rep2:
        type: str
      same:
        type: bool
      prob:
        type: num
        optional: true  

(untested, but it's something like this).

[ayonsn017]

Those are the two types right now which can be solved by an optional argument. In future, we expect a few more types of inputs which will not be similar to these two.

[stsievert]

we expect a few more types of inputs which will not be similar to these two.

We can specify that we expect one of several input types, or we can introduce other keys to expect all input types. We can find a workaround for that when it comes up – this issue isn't blocking.

[ayonsn017]

Changed the input from string to a list of dictionaries.